AQUAMan: QoE-driven cost-aware mechanism for SaaS acceptability rate adaptation

Abstract : As more interactive and multimedia-rich applications are migrating to the cloud, end-user satisfaction and her Quality of Experience (QoE) will become a determinant factor to secure success for any Software as a Service (SaaS) provider. Yet, in order to survive in this competitive market, SaaS providers also need to maximize their Quality of Business (QoBiz) and minimize costs paid to cloud providers. However, most of the existing works in the literature adopt a provider-centric approach where the end-user preferences are overlooked. In this article, we propose the AQUAMan mechanism that gives the provider a fine-grained QoE-driven control over the service acceptability rate while taking into account both end-users' satisfaction and provider's QoBiz. The proposed solution is implemented using a multi-agent simulation environment. The results show that the SaaS provider is capable of attaining the predefined acceptability rate while respecting the imposed average cost per user. Furthermore, the results help the SaaS provider identify the limits of the adaptation mechanism and estimate the best average cost to be invested per user.
Type de document :
Communication dans un congrès
International Conference on Web Intelligence, WI '17 , Aug 2017, Leipzig, Germany. Springer International Publishing, pp.Pages 331-339 Proceedings of the International Conference on Web Intelligence. 〈10.1145/3106426.3106485〉
Liste complète des métadonnées

https://hal-emse.ccsd.cnrs.fr/emse-01575677
Contributeur : Florent Breuil <>
Soumis le : lundi 21 août 2017 - 14:29:04
Dernière modification le : lundi 28 mai 2018 - 13:38:02

Identifiants

Citation

Amro Najjar, Yazan Mualla, Olivier Boissier, Gauthier Picard. AQUAMan: QoE-driven cost-aware mechanism for SaaS acceptability rate adaptation. International Conference on Web Intelligence, WI '17 , Aug 2017, Leipzig, Germany. Springer International Publishing, pp.Pages 331-339 Proceedings of the International Conference on Web Intelligence. 〈10.1145/3106426.3106485〉. 〈emse-01575677〉

Partager

Métriques

Consultations de la notice

194